Re: [Bacula-users] Restore dies every time

[Uwe S. <uwe...@ni...>]

On Fri, Jun 15, 2012 at 10:39:45AM +0200, Tilman Schmidt wrote:
> 
> 
> Am 13.06.2012 12:13, schrieb "M. Müller":
> > sometimes the restore seems to work and ends with Status=OK, but it 
> > allways dies. If Status is "OK", then files are restored, and hopefully 
> > all are restored.
> > The messages in the log file is allways: "bacula-fd: Bacula interrupted 
> > by signal 11: Segmentation violation"
> > bacula informs me via mail:
> [snip]
> > Bacula 5.2.6 and now 5.2.9 is compiled under OpenSuSE 12.1 with
> [...]
> > The backup up Client was Windows 2008 x64 with bacula 5.2.6 X64
> > 
> > I am a bit desperate as i surely need some files from that backup ;-)
> > Any hints are very much appreciated.
> 
> I observe similar behaviour with Bacula 5.2.* on CentOS 6, also with
> a Windows client. (2003 x86 in my case)
> 
> Restores to the Windows client systematically crash the FD on the
> client without restoring anything. This seems to be a known, as
> yet unsolved problem. There are several posts on this on the list.
> 
> Restoring the same files to the Bacula server instead works inasfar
> as the files are restored correctly, but crashes the FD on the
> server in the end, leaving a backtrace file in the Bacula working
> directory:
> 

I can confirm this behaviour on our setup (CentOS 6.2, latest patches,
bacula 5.2.6 compiled from source). 

Following a restore from a windows 2008 x64 R2 client, I have to
restart bacula-fd on the server. 

Attempt to dump current JCRs. njcrs=1
threadid=(nil) JobId=29132 JobStatus=T jcr=0x7f14980008e8
name=RestoreFiles.2012-06-15_11.07.26_56
threadid=(nil) killable=0 JobId=29132 JobStatus=T jcr=0x7f14980008e8
name=RestoreFiles.2012-06-15_11.07.26_56
	use_count=1
	JobType=R JobLevel= 
	sched_time=15-Jun-2012 11:07 start_time=15-Jun-2012 11:07
	end_time=01-Jan-1970 01:00 wait_time=01-Jan-1970 01:00
	db=(nil) db_batch=(nil) batch_started=0
Attempt to dump plugins. Hook count=0

All the best, Uwe